Software Deployment on Heterogeneous Platforms: A Systematic Mapping Study
收藏ieee-dataport.org2025-01-15 收录
下载链接:
https://ieee-dataport.org/documents/software-deployment-heterogeneous-platforms-systematic-mapping-study
下载链接
链接失效反馈官方服务:
资源简介:
Context: Multiple types of processing units (e.g., CPUs, GPUs and FPGAs) can be used jointly to achieve better performance in computational systems. However, these units are built with fundamentally different characteristics and demand attention especially towards software deployment. Objective: The goal of this work is to summarize the state-of-the art of software deployment on heterogeneous platforms. We provide an overview of the research area by searching for and categorizing relevant studies, as well as discussing gaps and trends of the field. We are interested in the main concerns (RQ1) and the approaches used (RQ2) when deploying software on heterogeneous platforms. Method: In order to achieve our goal, we performed a systematic mapping study, which refers to a method for reviewing literature with basis on predefined search strategies and a multi-step selection process. Results: We selected and analyzed 146 primary studies from multiple sources, and found that the area of research is dominated by solution proposals. The majority of the studies discussed concerns about scheduling, the quality of the software, and its architecture. A large number of studies focused on the problem of scheduling tasks and processes. We found approaches that are applied at different binding times (i.e., design time, runtime, orthogonal). Conclusion: The evaluation of the proposed solutions in an industrial context are missing. Also, the proposed methods have not been evaluated in development processes. Most of the methods address a particular concern, or a few concerns, while there is a lack of a holistic approach.
背景:在计算系统中,多种类型的处理单元(例如,CPU、GPU 和 FPGA)可以联合使用以实现更好的性能。然而,这些单元基于根本不同的特性,尤其在软件部署方面需要特别注意。目标:本工作的目标是总结异构平台上软件部署的最新进展。我们通过搜索和分类相关研究,以及对该领域的空白和趋势进行讨论,对研究领域进行了概述。我们关注的主要问题是(RQ1)在异构平台上部署软件时采用的方法(RQ2)。方法:为了实现我们的目标,我们进行了一项系统映射研究,这是一种基于预定义的搜索策略和多步选择过程的文献综述方法。结果:我们从多个来源选择了并分析了 146 项原始研究,发现该研究领域主要由解决方案提案主导。大多数研究讨论了关于调度、软件质量和其架构的问题。大量研究集中在调度任务和流程的问题上。我们发现了一些在不同绑定时间(即设计时间、运行时、正交)应用的方法。结论:在工业环境中对所提解决方案的评估尚显不足。此外,所提方法在开发过程中的评估也尚未进行。大多数方法针对特定的关注点,或者少数关注点,而缺乏一种整体的方法。
提供机构:
IEEE Dataport



